raster_aggregate: Aggregate multiple BlueSky model runs into a single...

In MazamaScience/AirFireModeling: Utilities for Working with Monitoring Data and Smoke Model Output

Description

Combines multiple Bluesky model runs to produce a single RasterBrick object. Users can set the first run, the last run, the hourly spacing between model runs, and the chunk index for the section of each model's time axis to be used.

Usage

raster_aggregate(
  modelName = NULL,
  firstModelRun = NULL,
  lastModelRun = NULL,
  modelMode = "forecast",
  baseUrl = "https://haze.airfire.org/bluesky-daily/output/standard",
  xlim = NULL,
  ylim = NULL,
  clean = TRUE,
  verbose = TRUE,
  spacing = NULL,
  chunk = 1
)

Arguments

modelName	Model identifier(s).
firstModelRun	Initialization datestamp of first model run as "YYYYmmddHH".
lastModelRun	Initialization datestamp of last model run as "YYYYmmddHH".
modelMode	Subdirectory path containing BlueSky output, i.e. 'forcast'.
baseUrl	Base URL for BlueSky output.
xlim	A vector of coordinate longitude bounds.
ylim	A vector of coordinate latitude bounds.
clean	Logical specifying removal of original model data after conversion to "v2" format.
verbose	Logical to display messages.
spacing	Time difference (hrs) between each model run – overrides default values.
chunk	The portion of the time (of width = 'spacing') to be chosen from each model run. If spacing = 12, then chunk = 1 corresponds to the first 12 hours of the each model run, chunk = 2 corresponds to the next 12 hours of each model run, and so on. See the note below for further explanation.

Details

Setting chunk = 1 results in grabbing the initial portion of each model run. Setting chunk to higher numbers utilizes times further out in the the forecast of each model run implying more uncertainty.

Value

A RasterBrick object.

Chunks

Suppose a model is run every 12 hours. For this model spacing = 12. In this case, chunk = 1 means we select the first 12 hours from each model run – i.e. we work with the model data from as close to initialization as possible. Setting chunk = 2 means we select the next 12 hours (i.e. hours 13-24) from each model run. Setting chunk = 3 means we select hours 25-48 from each model run and so on.

Chunking can be used to account for any inaccuracies in the data that may be due to poor initial conditions. For example, older models only predicted impacts from existing sources, and did not take into account any smoke that may have already been present at the time of initialization. In this case, setting chunk = 2 may produce better results.

Examples

library(AirFireModeling)
setModelDataDir('~/Data/BlueSky')

# Kincade fire
rasterBrick <- raster_aggregate(
  modelName = "PNW-4km",
  firstModelRun = 2020091300,
  lastModelRun = 2020091700,
  xlim = c(-123.5, -121.5),
  ylim = c(44.0, 46.0)
)

raster_facet(
  rasterBrick,
  title = "Oregon Labor Day Fires -- 2020091300 through 2020091700",
  ncol = 12,
  palette = 'Spectral',
  col_county = 'gray95',
  direction = -1,
  breaks = c(-Inf, 0, 12, 35, 55, 150, 250, 350, Inf)
)

MazamaScience/AirFireModeling documentation built on March 13, 2021, 12:02 a.m.